Class ResponseCreateParams.PromptCacheRetention

  • All Implemented Interfaces:
    com.openai.core.Enum

    
    public final class ResponseCreateParams.PromptCacheRetention
     implements Enum
                        

    The retention policy for the prompt cache. Set to 24h to enable extended prompt caching, which keeps cached prefixes active for longer, up to a maximum of 24 hours. Learn more. For gpt-5.5, gpt-5.5-pro, and future models, only 24h is supported.

    For older models that support both in_memory and 24h, the default depends on your organization's data retention policy:

    • Organizations without ZDR enabled default to 24h.

    • Organizations with ZDR enabled default to in_memory when prompt_cache_retention is not specified.